基于嵌入置信度的眼底血管边缘检测方法

Fundus Vessel Edge Detection Method Based on Embedded Confidence

针对现有的眼底图像血管检测方法对小血管分割效果差、计算量大以及自动化程度不高等问题,提出基于嵌入置信度的眼底血管边缘检测方法。改进传统的基于梯度的边缘检测算法,利用边缘置信度来辅助检测。实验结果表明,通过适当的参数选择,该方法能快速准确地检测出血管边缘,甚至包括弱小边缘,并能较好地抑制噪声的影响。

Embedded confidence fundus vessel detection method is proposed in terms of the bad effection, large calculation and low degree of automation of existing fundus vessel detection method. This method improves the traditional gradient-based edge detection algorithm, using the confidence of edge to assist detection. Experimental results show that by properly choosing the parameters, this method can rapidly and accurately detect the edge of the blood vessels, including small edge, and be able to inhibit the effects of noise.